Use of a Combination of Isotopically Coded Cross-Linkers and

Anal. Chem. 2010, 82, 817–823

Use of a Combination of Isotopically Coded Cross-Linkers and Isotopically Coded N-Terminal Modification Reagents for Selective Identification of Inter-peptide Crosslinks Evgeniy V. Petrotchenko, Jason J. Serpa, and Christoph H. Borchers* University of VictoriasGenome British Columbia Protein Centre, Department of Biochemistry and Microbiology, University of Victoria, No. 3101-4464 Markham Street, Vancouver Island Technology Park, Victoria, British Columbia V8Z7X8, Canada Cross-linking combined with mass spectrometry has great potential for determining three-dimensional structures of proteins and protein assemblies. One of the main analytical challenges of this method is the specific detection and identification of the inter-peptide crosslinks in the peptide mixture after enzymatic digestion of the cross-linked protein complex. These inter-peptide crosslinks are important because they provide the critical distance information needed for structural proteomics studies. In this paper, we demonstrate the use of isotopically coded N-terminal modification (ICNTM) in combination with isotopically coded cross-linkers (ICCL) for specific detection of inter-peptide crosslinks. Inter-peptide crosslinks contain two amino termini, compared to one in the case of free peptides, dead-end crosslinks, or intra-peptide crosslinks. Therefore, N-terminal modification with a 1:1 mixture of heavy and light isotopically coded reagents produces inter-peptide crosslinks with a distinct isotopic signature (a 1:2:1 ratio). Modification also occurs at the ε-amino groups of non-cross-linked lysine residues, resulting in two modifications per free lysine-containing peptide. However, if ICCL and ICNTM are used together, inter-peptide crosslinks can be distinguished from free lysine-containing peptides. Specialized software has also been developed for the analysis of ICCL + ICNTM experimental data. This procedure, combined with software for data analysis, provides a simple and rapid method for specific detection of inter-peptide crosslinks. The use of chemical cross-linking combined with mass spectrometry is a rapidly developing technique for structural proteomics.1,2 The distance constraints between the cross-linked amino acids, which can be obtained from this technique, are valuable for use in various aspects of protein structural studies, such as modeling three-dimensional structures of single proteins, detecting conformational changes, elucidating protein interaction interfaces, determining the topology of multiprotein assemblies, * To whom correspondence should be addressed. E-mail: christoph@ proteincentre.com. (1) Back, J. W.; de Jong, L.; Muijsers, A. O.; de Koster, C. G. J. Mol. Biol. 2003, 331, 303–313. (2) Sinz, A. Mass Spectrom. Rev. 2006, 25, 663–682. 10.1021/ac901637v  2010 American Chemical Society Published on Web 01/05/2010

and identifying protein interaction partners. The typical experimental design for the identification of cross-linked sites is to perform an enzymatic digestion on the cross-linked proteins, with subsequent mass spectrometric analysis of the digest in order to find the cross-linked peptides (crosslinks). A lot of progress has been made recently in terms of the development of new reagents,3 the use of high mass-accuracy instrumentation,4 and the application of specialized data-analysis software5 designed especially for cross-linking experiments. One of the current bottlenecks of this approach is the detection and characterization of the inter-peptide crosslinks. Unfortunately, the inter-peptide crosslinksswhich provide the distance information between the cross-linked sitessare usually overwhelmed in the mass spectra by free peptides and less-informative dead-end and intra-peptide crosslinks. Selective identification and, if possible, isolation of the inter-peptide crosslinks would bring this method to the next level, allowing researchers to successfully tackle high-complexity protein systems and proteome-wide structural proteomics applications. Several solutions have been proposed for the selective detection of the inter-peptide crosslinks. Seebacher et al. and Chu et al. proposed differential incorporation of 16O/18O water during hydrolysis of the dead-end crosslinks.6,7 Conducting the crosslinking reaction in H216O or H218O results in a 2 Da isotopic signature for dead-end crosslinks but not for intra- and interpeptide crosslinks. Isotopically coded cleavable cross-linkers can also be used to distinguish between crosslink types.8,9 Cleavage of the cross-linker leads to the appearance of new isotopic signatures for fragments of cleaved peptide crosslinks due to the loss of part of the isotopic label. The cleaved fragments are related by characteristic mass differences from the masses of the uncleaved peptide crosslinks. These differences are specific (3) Petrotchenko, E. V.; Borchers, C. H. Mass Spectrom. Rev., submitted for publication. (4) Kalkhof, S.; Ihling, C.; Mechtler, K.; Sinz, A. Anal. Chem. 2005, 77, 495– 503. (5) Lee, J. Y. Mol. Biosyst. 2008, 4, 816–823. (6) Seebacher, J.; Mallick, P.; Zhang, N.; Eddes, J. S.; Aebersold, R.; Gelb, M. H. J. Proteome Res. 2006, 5, 2270–2282. (7) Chu, F.; Mahrus, S.; Craik, C. S.; Burlingame, A. L. J. Am. Chem. Soc. 2006, 128, 10362–10363. (8) Petrotchenko, E. V.; Olkhovik, V. K.; Borchers, C. H. Mol. Cell. Proteomics 2005, 4, 1167–1179. (9) Petrotchenko, E. V.; Xiao, K.; Cable, J.; Chen, Y.; Dokholyan, N. V.; Borchers, C. H. Mol. Cell. Proteomics 2009, 8, 273–286.

Analytical Chemistry, Vol. 82, No. 3, February 1, 2010

817

Figure 1. Principle of selective detection of inter-peptide crosslinks using N-terminal modification with isotopically coded reagents (ICNTM). N-Terminal modification of the inter-peptide crosslink (lower panel) produces distinct a isotopic signature of peaks with a characteristic 1:2:1 intensity distribution (ref 11): K, lysine residue; open circle, PCAS modification; OH, dead-end cross-link; solid line connecting K residues, cross-linker.

for each crosslink type (dead-end, intra-peptide, or inter-peptide crosslink). By employing the mass relationships between the cleaved and uncleaved crosslinks, the type of crosslink can be readily determined. Back et al. used proteolytic enzyme-catalyzed incorporation of 16 O/18O into the C-termini of peptides for distinguishing of the inter-peptide crosslinks.10 Because inter-peptide crosslinks possess two C-termini compared to one in the case of free peptides, dead-end, or intra-peptide crosslinks, they can incorporate two isotopic labels, thus displaying a specific isotopic signature in the mass spectra. Chen et al. modified the N-termini of peptides from enzymatic digests of cross-linked proteins with amino-reactive isotopically coded reagent mixtures.11 Because inter-peptide crosslinks possess two amino termini, this labeling results in two modifications per cross-link. When a 1:1 mixture of light and heavy isotopic forms of the modifying reagent is used, crosslinks exhibit a specific isotopic signature with a 1:2:1 intensity ratio, whereas free peptides, dead-end, and intra-peptide crosslinks exhibit a 1:1 isotopic pattern. Unfortunately, any peptide that contains a lysine residue with a free amino group also results in two modifications and exhibits this same 1:2:1 ratio, producing a misidentification as a cross-link. To avoid this, the authors suggested blocking any free amino groups on the cross-linked protein by methylation prior to enzymatic digestion and isotopically coded N-terminal modification (ICNTM). This leads to a rather complicated multistep labeling procedure and, theoretically, could lead to higher crosslink molecular weights due to missed tryptic cleavage at the blocked lysines. In this paper, we propose a revision of the original method by using isotopically coded cross-linking (ICCL) and ICNTM together. Using two different reagents for cross-linking (10) Back, J. W.; Notenboom, V.; de Koning, L. J.; Muijsers, A. O.; Sixma, T. K.; de Koster, C. G.; de Jong, L. Anal. Chem. 2002, 74, 4417–4422. (11) Chen, X.; Chen, Y. H.; Anderson, V. E. Anal. Biochem. 1999, 273, 192– 203.

818


and N-terminal modification, with distinct isotopic codings, allows us to differentiate inter-peptide crosslinks from free lysinecontaining peptides without the need for blocking free amino groups in the proteins. Incorporating a search for the specific isotopic signatures into the software algorithm results in a simplified procedure for the rapid and specific detection of interpeptide crosslinks in the mass spectra. EXPERIMENTAL SECTION Materials. All materials were from Sigma-Aldrich, unless otherwise noted. Pyridine carboxylic acid succinimide (PCAS) was synthesized from 3-pyridinecarboxylic acid by activation with N-hydroxysuccinimide in presence of dicyclohexylcarbodiimide. Light and heavy forms of the reagent, PCAS-H4 and PCAS-D4, were obtained by using H4- and D4-3-pyridinecarboxylic acid (C/D/N Isotopes), respectively. Cross-Linking and N-Terminal Modification Analysis of Human Immunodeficiency Virus Reverse Transcriptase. A 10 µL aliquot of a 1 mg/mL solution of human immunodeficiency virus reverse transcriptase (HIV-RT; sequence accession numbers GI:1827604 and 1827605) (Worthington Biochemical Corp.) in phosphate-buffered saline was mixed with 1 µL of a 0.5 mM bis(sulfosuccinimidylsuberate-H12/D12) (BS3-H12/D12) (Creative Molecules, Inc.) solution in water, prepared from 50 mM stock solution of the cross-linker. The pH of the mixture was adjusted to 8.0-8.5 by the addition of 0.2 M Na2HPO4. The reaction mixture was incubated for 30 min at 25 °C. The mixture was then digested with sequencing-grade trypsin (Promega) overnight at 25 °C, at a 10:1 substrate/enzyme ratio. The resulting peptide mixture was split in halfsone-half was modified with ICNTM while the other half was not. For the ICNTM fraction, PCAS-H4/D4 was added to give a final molar ratio of 1:10 protein/PCAS. The reaction was allowed to proceed for 2.5 h at 25 °C. Ammonium hydroxide was added to both samples to

Figure 2. Schematic representation of the peak signals in the mass spectra of different peptide species when using isotopically coded crosslinkers (ICCL), isotopically coded N-terminal modification reagents (ICNTM), and a combination of the both reagents (ICCL + ICNTM) are used. Symbols used are the same as in Figure 1, except that the single and double lines connecting the lysine residues represent non-isotopicallycoded cross-linkers and equimolar mixture of light and heavy isotopic forms of isotopically coded cross-linkers, respectively.

give a final concentration of 2 M, and the reaction mixtures were incubated for an additional 4 h at 25 °C. The samples were acidified with trifluoroacetic acid (TFA) and separated by nanoflow reversed-phase high-performance liquid chromatography (HPLC) on a 1D Tempo nano-LC system (Eksigent) equipped with an LC Packings 0.3 mm × 5 mm C18 PepMap guard column (5 µm particle size, 100 Å pore size) and a 75 µm × 15 cm capillary column packed in-house with Magic C18 Aq (Michrom Bioresources Inc.) particles (5 µm, 100 Å). This capillary LC system was operated at a flow rate of 300 nL/min, using a 55 min gradient from 5% to 60% acetonitrile (0.1% TFA). The column effluent was spotted at 1 min intervals (300 nL/spot) onto a stainless steel MALDI target using a Dionex Probot spotter. The spots were dried, overlaid with 0.5 µL 1 mg/mL R-cyano-4-hydroxycinnamic acid matrix solution,

and analyzed by matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) using an Applied Biosystems 4800 MALDI time-of-flight/time-of-flight (MALDI-TOF/TOF) mass spectrometer.12 MALDI-MS data was acquired automatically over a mass range of 800-4000 Da in the positive-ion reflector mode using a fixed laser intensity for 1000 shots/spectrum, with a uniformly random spot search pattern. Mass spectra were analyzed with the DX program in the ICC-CLASS software suite13 (www.creativemolecules.com), using 0.05 Da mass tolerance for the detection of the doublets and with our new specialized program for identification of ICCL + ICNTM products, which is described below. The proposed (12) Kuzyk, M. A.; Ohlund, L. B.; Elliott, M. H.; Smith, D.; Qian, H.; Delaney, A.; Hunter, C. L.; Borchers, C. H. Proteomics 2009, 9, 3328–3340. (13) Petrotchenko, E. V.; Borchers, C. H. BMC Bioinf., in press, 2009.


819

instrument-independent. All results are saved in text files as tabdelimited values and therefore can be easily copied into Excel spreadsheets.

Figure 3. Flowchart of the data analysis of ICCL + ICNTM experiments. MS peaklists of chromatographic runs from ICCL and ICCL + ICNTM experiments are searched for doublets of signals 12.07 Da apart and multiplets 4.03 Da apart by the DX and the DXMultipletFinder programs, respectively. Twelve-Da doublet and four-Da multiplet mass lists are filtered by the DXMassListFilter program to produce mass list of unique doublets and multiplets. DXDX ICCL+ICNTM program compares mass lists from ICCL and ICCL + ICNTM experiments and identifies matches between isotopically coded crosslinks and isotopically coded N-terminally modified crosslinks based on the mass of the ICNTM reagent (105.02 Da for PCAS-H4/ D4). Matches that are found are identified as inter-peptide crosslink candidates.

assignments of the inter-peptide crosslink candidates were confirmed by tandem mass spectrometry (MS/MS) analysis using the DXMSMS program in the ICC-CLASS software suite. A 1 kV MS/MS operating mode was used, the relative precursor mass window was set at 50 (full width at half-maximum), and CID using air was turned off, with metastable suppression enabled. Software for ICCL + ICNTM Data Analysis. The search for doublets of signals 12 Da apart was performed using the DX program of the ICC-CLASS software suite (Creative Molecules Inc.). The search for multiplets 4 Da apart was performed by DXMultipletFinder program. The program searches sets of peaklists of MS spectra from chromatographic fractions for a series of peaks separated by the specific mass difference corresponding to the isotopic coding of the NTM reagent. DX and DXMultipletFinder produce mass lists of doublets and multiplets, respectively. These mass lists can be filtered by the DXMassListFilter program to remove repeats for the doublets and multiplets in order to produce mass lists containing unique doublets and multiplets. The DXDX ICCL + ICNTM program compares mass lists from ICCL and ICCL + ICNTM experiments and identifies matches between isotopically coded crosslinks and isotopically coded N-terminally modified crosslinks, based on the mass of the ICNTM reagent (105.02 Da for PCAS-H4/D4). These software programs were written in Microsoft Visual Basic 2008 Express edition and are freely available from the www.creativemolecules.com Web site. Executing the downloaded programs requires installation of Microsoft NET Framework, which is available as a free download from www.microsoft.com. All of these programs are primarily oriented toward Applied Biosystems MALDI-TOF/TOF data on the HPLC fractions but can be used with any tab-separated mass lists and therefore are 820


RESULTS AND DISCUSSION ICCL + ICNTM. N-terminal modification of peptides with isotopically coded reagents for identification of inter-peptide crosslinks was first proposed by Chen et al.11 (Figure 1). The specific mass spectrometric signature for an inter-peptide crosslink is based on the presence of two N-terminal amino groups in an inter-peptide crosslink compared to one N-terminal amino acid in each un-cross-linked free peptide, dead-end, or intra-peptide crosslink. Modification of an inter-peptide crosslink with an equimolar mixture of light and heavy isotopic forms of an amine-reactive reagent results in a specific triplet of signals separated by mass according to the mass difference between the light and heavy isotopic forms of the reagent. The resulting 1:2:1 intensity ratio of these peaks is due to the possible combinations of the different isoforms of the product (LL, LH + HL, HH, where L and H are light and heavy forms, respectively). Modification of the single N-terminal amino group of free peptides, dead-end, and intrapeptide crosslinks results in a doublet of signals in a 1:1 ratio due to the possible combinations of L and H isoforms. Unfortunately, peptides containing lysine residues contain an additional, second, amino group which can also be modified and which can consequently produce a 1:2:1 triplet signature after modification with the reagent. Chen et al. originally performed an extra stepsblocking ε-amino groups of lysine residues by methylation prior to the enzymatic digestion of the cross-linked proteins.11 This additional chemical derivatization step complicates the procedure, makes it vulnerable to additional side reactions, and produces longer tryptic peptides which are generally more difficult to analyze. As an alternative, we propose the use of isotopically coded cross-linkers in combination with isotopically coded N-terminal modification reagents (ICCL + ICNTM). When cross-linked with an equimolar mixture of light and heavy isotopic forms of the cross-linking reagent, peptide crosslinks appear in the spectrum as doublets of peaks of equal intensity. The subsequent modification with isotopically coded amine-reactive reagent generates a combination of the products as described above for each light and heavy crosslink form (Figure 2). Superposition of these two isotopic mass spectrometric signatures from both cross-linker and modifying reagent allows one to discriminate inter-peptide crosslinks from free peptides, free lysine-containing peptides, and from the majority of the dead-end and intra-peptide crosslinks. Only relatively rare dead-end or intra-peptides crosslinks containing free lysine residues will produce isotopic signatures similar to interpeptide crosslink isotopic signatures. Data Analysis of the ICCL + ICNTM Experiments. To automate the identification of the mass spectrometric signatures specific to inter-peptide crosslinks, we have developed a specialized software program called DXDX ICCL+ICNTM. The search algorithm is based on the detection of the different doublets resulting from the isotopic coding of both cross-linking and modification reagents. For example, if cross-linker is labeled with 12 deuterium atoms and a modification reagent is labeled with 4 deuteriums, then the program will search for peaks which will satisfy both criteriashaving a 12 Da doublet and having a series

Figure 4. MS and MS/MS spectra of the inter-peptide crosslink, LKPGMDGPK-ELNKR, for the ICCL (top), ICNTM (middle), and ICCL + ICNTM (bottom) procedures. X ) H or D for equimolar mixtures of ICCL and ICNTM reagents. Table 1. ICCL + ICNTM Analysis of the HIV-Reverse Transcriptase Using BS3-H12/D12 and PCAS-H4/D4 as Isotopically Coded Cross-Linker and N-Terminal Modification Reagent, Respectivelya

a ICCL, experiment with cross-linking only; ICCL + ICNTM, experiment with cross-linking and N-terminal modification; CL, crosslink type; Mobs, Mtheor, [M + H]+, observed and theoretical masses of light isotopic form of the crosslink in Da; ∆, error in ppm; RT, retention time in min; signature, ICNTM isotopic signature; Rs, Re, start and end residue number of the peptide; sequence, peptide sequence (cross-linking sites are highlighted in bold); cross-link, schematic representation of the modified crosslinks, symbols are as in Figure 1.


821

Figure 5. Crystal structure of HIV-RT with annotated cross-linked residues. The nitrogen atoms of the ε-amino groups of the cross-linked lysine residues are shown as connected spheres. Residue numbers are annotated for the A and B chains according to PDB coordinates file (code 1DLO); distances were calculated using the coordinates of the NZ atoms of the lysine residues. Residues B287 and A220 are too far from B82 and A70, respectively, to form crosslinks, which indicates that the detected 82-287 and 70-220 crosslinks must belong to chains A and B, respectively. These crosslinks allow one to distinguish between the two distinct conformational states of the A and B chains.

of 4 Da doublets (multiplet) (see Figure 2, right column, four bottom panels). A flowchart of the program is shown in Figure 3. The search can be performed on a single ICCL + ICNTM chromatogram as well as on two separate runs: ICCL alone and ICCL + ICNTM. The program produces a list of the detected masses of the inter-peptide crosslink candidates as tab-separated values in a text file. ICCL + ICNTM Analysis of HIV-RT. To demonstrate the utility of this approach, we applied it to a study of the model heterodimeric protein, HIV-RT. HIV-RT was cross-linked with isotopically coded BS3-H12/D12 and digested with trypsin. The sample was divided into two parts: one-half was modified with PCAS-H4/D4, and both samples were treated with ammonia. The resulting peptide mixtures were separated by HPLC and were analyzed by MALDI-MS. Spectra from the chromatographic fractions were screened for doublets of signals 12.07 Da apart for the sample without PCAS modification (ICCL) and for series of doublets of signals 4.05 Da apart (multiplets) for the sample with PCAS modification (ICCL + ICNTM). Both mass lists were used as input for the inter-peptide crosslinks identification program, DXDX ICCL+ICNTM. Peaks corresponding to possible inter-peptide crosslinks were confirmed by MS/MS analysis (Figure 4). 822


A search of the ICCL 1000-4000 Da mass spectra data set with DX using 0.05 Da tolerance for doublets detection and 150 counts for peak intensity cutoff produced 263 12 Da doublets. Filtering out repeats with DXMassListFilter produced 169 unique 12 Da doublets. Search of the ICCL + ICNTM data set with DXMultipletFinder produced 285 4 Da multiplets. Filtering with DXMassListFilter resulted in 200 unique multiplets. Using these two mass lists as an input to the DXDX ICCL+ICNTM program resulted in the identification of 11 inter-peptide crosslink candidates, 7 of which were confirmed by MS/MS analysis and assigned with DXMSMS program as Lys-Lys tryptic crosslinks. Four of these were confirmed as inter-peptide crosslinks, two were found to be dead-ends, and one was an intra-peptide crosslink containing free lysine residues (Table 1). All of the inter-peptide crosslinks were in a good agreement with the crystal structure (PDB code: 1DLO) and were within the maximum span of the BS3 cross-linker (11 A). These inter-peptide crosslinks allow us to distinguish between the different conformational states of the A and B chains of HIV-RT (Figure 5). Using the combination of ICCL + ICNTM significantly reduced the number of potential candidates from hundreds to only 11, thus eliminating the need to analyze hundreds of crosslink peaks by MS/MS and

allowing us to focus our analysis on only a few possible interpeptide crosslinks. Treatment with ammonia appears to be a necessary step in this procedure, as it reverses both the cross-linker and the NTM reagent modifications of the hydroxyl groups on the peptides. Using ICNTM alone (using a noncoded form of BS3), with and without ammonia treatment, resulted in 314 and 112 4 Da multiplets, respectively. Narrowing down the cross-linking reaction products to only modified amino groups, by base hydrolysis of the O-acyl derivatives of serine, threonine, and tyrosine, is usually desirable. We did not perform the entire procedure without ammonia treatment, but one would expect a similar factor of 3 decrease in the number of ICNTM-modified peptides after ammonia treatment because the reactive groups of the cross-linker reagent and of the ICNTM reagent are both NHS esters. An additional benefit of the procedure, as was noted in the original ICNTM work by Chen et al., is the formation of specific isotopic signatures for the fragment masses of the inter-peptide crosslinks.11 The combination of two different isotopic codingss from the cross-linker and from the modification reagentsassists in the assignment of the fragment ions of the inter-peptide crosslinks. For example, bCL- and yCL-ions (ions which contain fragments from one peptide, the cross-linker bridge, and the intact attached peptide) will show a combination of the isotopic signatures from both the cross-linker and ICNTM reagent. In contrast, the b-ions from the non-cross-linker-containing fragments of one of the peptides will always appear as 4 Da doublets, whereas the y-ions will be labeled or unlabeled, depending on presence or absence of a C-terminal lysine (Figure 4). Using the 120 kDa HIV-RT as an example, we successfully applied our ICCL + ICNTM procedure to the selective detection

and identification of inter-peptide crosslinks in a protein system of reasonable complexity. Identification by MALDI-MS of four inter-peptide crosslinks which are compatible with the known crystal structure of the protein complex has provided satisfactory validation of this approach. Utilizing both reagents in isotopically coded form allowed us to eliminate the need for chemical blocking of the lysine residues in the intact cross-linked protein and resulted in a simplified one-step modification procedure for selective identification of the inter-peptide crosslinks. Together with the new specialized program for data analysis, this procedure is a rapid method for the specific identification of inter-peptide crosslinks. CONCLUSIONS We report here a variation of the method of N-terminal modification of the peptides with isotopically coded reagents for the specific detection of the inter-peptide crosslinks. Instead of blocking lysine residues prior to digestion of the cross-linked proteins, we use isotopically coded cross-linkers. Utilizing both cross-linker and modifying reagent in isotopic coding forms produces specific isotopic signatures for inter-peptide crosslinks. When combined with a program for the screening of the mass spectra for these specific isotopic signatures, this provides a method for the rapid discrimination of inter-peptide crosslinks from the overwhelming pool of free peptides, dead-end, and intrapeptide crosslinks. ACKNOWLEDGMENT We thank Genome Canada and Genome BC for Technology Development grant and platform funding and Dr. Carol Parker for critical reading of the manuscript. Received for review July 22, 2009. Accepted November 23, 2009. AC901637V


823

Use of a Combination of Isotopically Coded Cross-Linkers and

Recommend Documents